Home Explore Help
Register Sign In
publics
/
net
0
0
Fork 0
Files Issues 0 Pull Requests 0 Wiki
Browse Source

html: implement generic raw text element parsing algorithm

See: https://html.spec.whatwg.org/multipage/parsing.html#parsing-elements-that-contain-only-text

This follows up on golang.org/cl/205617

Change-Id: Id99054bc25e9ea90bb3f03b15c14c13573520997
Reviewed-on: https://go-review.googlesource.com/c/net/+/210318
Run-TryBot: Kunpei Sakai <namusyaka@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Nigel Tao <nigeltao@golang.org>
Kunpei Sakai 6 years ago
parent
afd1edf42a
commit
1ddd1de85c
5 changed files with 332 additions and 150 deletions
Split View Show Diff Stats
  1. 33 15
      html/parse.go
  2. 2 1
      html/parse_test.go
  3. 197 116
      html/testdata/webkit/tests16.dat
  4. 37 18
      html/testdata/webkit/tests5.dat
  5. 63 0
      html/testdata/webkit/webkit02.dat

+ 33 - 15
html/parse.go
View File 

@@ -184,6 +184,17 @@ func (p *parser) clearStackToContext(s scope) {
 
 	}
 
 }
 
 
+// parseGenericRawTextElements implements the generic raw text element parsing
 
+// algorithm defined in 12.2.6.2.
 
+// https://html.spec.whatwg.org/multipage/parsing.html#parsing-elements-that-contain-only-text
 
+// TODO: Since both RAWTEXT and RCDATA states are treated as tokenizer's part
 
+// officially, need to make tokenizer consider both states.
 
+func (p *parser) parseGenericRawTextElement() {
 
+	p.addElement()
 
+	p.originalIM = p.im
 
+	p.im = textIM
 
+}
 
+
 
 // generateImpliedEndTags pops nodes off the stack of open elements as long as
 
 // the top node has a tag name of dd, dt, li, optgroup, option, p, rb, rp, rt or rtc.
 
 // If exceptions are specified, nodes with that name will not be popped off.
 
@@ -631,19 +642,23 @@ func inHeadIM(p *parser) bool {
 
 			p.acknowledgeSelfClosingTag()
 
 			return true
 
 		case a.Noscript:
 
-			p.addElement()
 
 			if p.scripting {
 
-				p.setOriginalIM()
 
-				p.im = textIM
 
-			} else {
 
-				p.im = inHeadNoscriptIM
 
+				p.parseGenericRawTextElement()
 
+				return true
 
 			}
 
+			p.addElement()
 
+			p.im = inHeadNoscriptIM
 
+			// Don't let the tokenizer go into raw text mode when scripting is disabled.
 
+			p.tokenizer.NextIsNotRawText()
 
 			return true
 
-		case a.Script, a.Title, a.Noframes, a.Style:
 
+		case a.Script, a.Title:
 
 			p.addElement()
 
 			p.setOriginalIM()
 
 			p.im = textIM
 
 			return true
 
+		case a.Noframes, a.Style:
 
+			p.parseGenericRawTextElement()
 
+			return true
 
 		case a.Head:
 
 			// Ignore the token.
 
 			return true
 
@@ -1023,18 +1038,21 @@ func inBodyIM(p *parser) bool {
 
 			p.popUntil(buttonScope, a.P)
 
 			p.reconstructActiveFormattingElements()
 
 			p.framesetOK = false
 
-			p.addElement()
 
-			p.setOriginalIM()
 
-			p.im = textIM
 
+			p.parseGenericRawTextElement()
 
 		case a.Iframe:
 
 			p.framesetOK = false
 
+			p.parseGenericRawTextElement()
 
+		case a.Noembed:
 
+			p.parseGenericRawTextElement()
 
+		case a.Noscript:
 
+			if p.scripting {
 
+				p.parseGenericRawTextElement()
 
+				return true
 
+			}
 
+			p.reconstructActiveFormattingElements()
 
 			p.addElement()
 
-			p.setOriginalIM()
 
-			p.im = textIM
 
-		case a.Noembed, a.Noscript:
 
-			p.addElement()
 
-			p.setOriginalIM()
 
-			p.im = textIM
 
+			// Don't let the tokenizer go into raw text mode when scripting is disabled.
 
+			p.tokenizer.NextIsNotRawText()
 
 		case a.Select:
 
 			p.reconstructActiveFormattingElements()
 
 			p.addElement()

+ 2 - 1
html/parse_test.go
View File 

@@ -272,7 +272,8 @@ func TestParserWithoutScripting(t *testing.T) {
 
 |   <head>
 
 |     <noscript>
 
 |   <body>
 
-|     "<img src='https://golang.org/doc/gopher/frontpage.png' />"
 
+|     <img>
 
+|       src="https://golang.org/doc/gopher/frontpage.png"
 
 |     <p>
 
 |       <img>
 
 |         src="https://golang.org/doc/gopher/doc.png"

File diff suppressed because it is too large
+ 197 - 116
html/testdata/webkit/tests16.dat


+ 37 - 18
html/testdata/webkit/tests5.dat
View File 

@@ -1,8 +1,7 @@
 
 #data
 
 <style> <!-- </style>x
 
 #errors
 
-Line: 1 Col: 7 Unexpected start tag (style). Expected DOCTYPE.
 
-Line: 1 Col: 22 Unexpected end of file. Expected end tag (style).
 
+(1,7): expected-doctype-but-got-start-tag
 
 #document
 
 | <html>
 
 |   <head>
 
@@ -14,7 +13,8 @@ Line: 1 Col: 22 Unexpected end of file. Expected end tag (style).
 
 #data
 
 <style> <!-- </style> --> </style>x
 
 #errors
 
-Line: 1 Col: 7 Unexpected start tag (style). Expected DOCTYPE.
 
+(1,7): expected-doctype-but-got-start-tag
 
+(1,34): unexpected-end-tag
 
 #document
 
 | <html>
 
 |   <head>
 
@@ -27,7 +27,7 @@ Line: 1 Col: 7 Unexpected start tag (style). Expected DOCTYPE.
 
 #data
 
 <style> <!--> </style>x
 
 #errors
 
-Line: 1 Col: 7 Unexpected start tag (style). Expected DOCTYPE.
 
+(1,7): expected-doctype-but-got-start-tag
 
 #document
 
 | <html>
 
 |   <head>
 
@@ -39,7 +39,7 @@ Line: 1 Col: 7 Unexpected start tag (style). Expected DOCTYPE.
 
 #data
 
 <style> <!---> </style>x
 
 #errors
 
-Line: 1 Col: 7 Unexpected start tag (style). Expected DOCTYPE.
 
+(1,7): expected-doctype-but-got-start-tag
 
 #document
 
 | <html>
 
 |   <head>
 
@@ -51,7 +51,7 @@ Line: 1 Col: 7 Unexpected start tag (style). Expected DOCTYPE.
 
 #data
 
 <iframe> <!---> </iframe>x
 
 #errors
 
-Line: 1 Col: 8 Unexpected start tag (iframe). Expected DOCTYPE.
 
+(1,8): expected-doctype-but-got-start-tag
 
 #document
 
 | <html>
 
 |   <head>
 
@@ -63,7 +63,9 @@ Line: 1 Col: 8 Unexpected start tag (iframe). Expected DOCTYPE.
 
 #data
 
 <iframe> <!--- </iframe>->x</iframe> --> </iframe>x
 
 #errors
 
-Line: 1 Col: 8 Unexpected start tag (iframe). Expected DOCTYPE.
 
+(1,8): expected-doctype-but-got-start-tag
 
+(1,36): unexpected-end-tag
 
+(1,50): unexpected-end-tag
 
 #document
 
 | <html>
 
 |   <head>
 
@@ -75,7 +77,8 @@ Line: 1 Col: 8 Unexpected start tag (iframe). Expected DOCTYPE.
 
 #data
 
 <script> <!-- </script> --> </script>x
 
 #errors
 
-Line: 1 Col: 8 Unexpected start tag (script). Expected DOCTYPE.
 
+(1,8): expected-doctype-but-got-start-tag
 
+(1,37): unexpected-end-tag
 
 #document
 
 | <html>
 
 |   <head>
 
@@ -88,7 +91,8 @@ Line: 1 Col: 8 Unexpected start tag (script). Expected DOCTYPE.
 
 #data
 
 <title> <!-- </title> --> </title>x
 
 #errors
 
-Line: 1 Col: 7 Unexpected start tag (title). Expected DOCTYPE.
 
+(1,7): expected-doctype-but-got-start-tag
 
+(1,34): unexpected-end-tag
 
 #document
 
 | <html>
 
 |   <head>
 
@@ -101,7 +105,9 @@ Line: 1 Col: 7 Unexpected start tag (title). Expected DOCTYPE.
 
 #data
 
 <textarea> <!--- </textarea>->x</textarea> --> </textarea>x
 
 #errors
 
-Line: 1 Col: 10 Unexpected start tag (textarea). Expected DOCTYPE.
 
+(1,10): expected-doctype-but-got-start-tag
 
+(1,42): unexpected-end-tag
 
+(1,58): unexpected-end-tag
 
 #document
 
 | <html>
 
 |   <head>
 
@@ -113,7 +119,7 @@ Line: 1 Col: 10 Unexpected start tag (textarea). Expected DOCTYPE.
 
 #data
 
 <style> <!</-- </style>x
 
 #errors
 
-Line: 1 Col: 7 Unexpected start tag (style). Expected DOCTYPE.
 
+(1,7): expected-doctype-but-got-start-tag
 
 #document
 
 | <html>
 
 |   <head>
 
@@ -125,7 +131,7 @@ Line: 1 Col: 7 Unexpected start tag (style). Expected DOCTYPE.
 
 #data
 
 <p><xmp></xmp>
 
 #errors
 
-XXX: Unknown
 
+(1,3): expected-doctype-but-got-start-tag
 
 #document
 
 | <html>
 
 |   <head>
 
@@ -136,7 +142,7 @@ XXX: Unknown
 
 #data
 
 <xmp> <!-- > --> </xmp>
 
 #errors
 
-Line: 1 Col: 5 Unexpected start tag (xmp). Expected DOCTYPE.
 
+(1,5): expected-doctype-but-got-start-tag
 
 #document
 
 | <html>
 
 |   <head>
 
@@ -147,7 +153,7 @@ Line: 1 Col: 5 Unexpected start tag (xmp). Expected DOCTYPE.
 
 #data
 
 <title>&amp;</title>
 
 #errors
 
-Line: 1 Col: 7 Unexpected start tag (title). Expected DOCTYPE.
 
+(1,7): expected-doctype-but-got-start-tag
 
 #document
 
 | <html>
 
 |   <head>
 
@@ -158,7 +164,7 @@ Line: 1 Col: 7 Unexpected start tag (title). Expected DOCTYPE.
 
 #data
 
 <title><!--&amp;--></title>
 
 #errors
 
-Line: 1 Col: 7 Unexpected start tag (title). Expected DOCTYPE.
 
+(1,7): expected-doctype-but-got-start-tag
 
 #document
 
 | <html>
 
 |   <head>
 
@@ -169,8 +175,7 @@ Line: 1 Col: 7 Unexpected start tag (title). Expected DOCTYPE.
 
 #data
 
 <title><!--</title>
 
 #errors
 
-Line: 1 Col: 7 Unexpected start tag (title). Expected DOCTYPE.
 
-Line: 1 Col: 19 Unexpected end of file. Expected end tag (title).
 
+(1,7): expected-doctype-but-got-start-tag
 
 #document
 
 | <html>
 
 |   <head>
 
@@ -181,7 +186,9 @@ Line: 1 Col: 19 Unexpected end of file. Expected end tag (title).
 
 #data
 
 <noscript><!--</noscript>--></noscript>
 
 #errors
 
-Line: 1 Col: 10 Unexpected start tag (noscript). Expected DOCTYPE.
 
+(1,10): expected-doctype-but-got-start-tag
 
+(1,39): unexpected-end-tag
 
+#script-on
 
 #document
 
 | <html>
 
 |   <head>
 
@@ -189,3 +196,15 @@ Line: 1 Col: 10 Unexpected start tag (noscript). Expected DOCTYPE.
 
 |       "<!--"
 
 |   <body>
 
 |     "-->"
 
+
 
+#data
 
+<noscript><!--</noscript>--></noscript>
 
+#errors
 
+(1,10): expected-doctype-but-got-start-tag
 
+#script-off
 
+#document
 
+| <html>
 
+|   <head>
 
+|     <noscript>
 
+|       <!-- </noscript> -->
 
+|   <body>

+ 63 - 0
html/testdata/webkit/webkit02.dat
View File 

@@ -26,6 +26,23 @@
 
 |       <span>
 
 |         "B"
 
 
+#data
 
+<p id="status"><noscript><strong>A</strong></noscript><span>B</span></p>
 
+#errors
 
+(1,15): expected-doctype-but-got-start-tag
 
+#script-off
 
+#document
 
+| <html>
 
+|   <head>
 
+|   <body>
 
+|     <p>
 
+|       id="status"
 
+|       <noscript>
 
+|         <strong>
 
+|           "A"
 
+|       <span>
 
+|         "B"
 
+
 
 #data
 
 <div><sarcasm><div></div></sarcasm></div>
 
 #errors
 
@@ -136,6 +153,52 @@
 
 |     <input>
 
 |     <table>
 
 
+#data
 
+<b><em><foo><foo><aside></b>
 
+#errors
 
+#document
 
+| <html>
 
+|   <head>
 
+|   <body>
 
+|     <b>
 
+|       <em>
 
+|         <foo>
 
+|           <foo>
 
+|     <em>
 
+|       <aside>
 
+|         <b>
 
+
 
+#data
 
+<b><em><foo><foo><aside></b></em>
 
+#errors
 
+#document
 
+| <html>
 
+|   <head>
 
+|   <body>
 
+|     <b>
 
+|       <em>
 
+|         <foo>
 
+|           <foo>
 
+|     <em>
 
+|     <aside>
 
+|       <em>
 
+|         <b>
 
+
 
+#data
 
+<b><em><foo><foo><foo><aside></b>
 
+#errors
 
+#document
 
+| <html>
 
+|   <head>
 
+|   <body>
 
+|     <b>
 
+|       <em>
 
+|         <foo>
 
+|           <foo>
 
+|             <foo>
 
+|     <aside>
 
+|       <b>
 
+
 
 #data
 
 <b><em><dcell><postfield><postfield><postfield><postfield><missing_glyph><missing_glyph><missing_glyph><missing_glyph><hkern><aside></b></em>
 
 #errors

Some files were not shown because too many files changed in this diff